commit 1432785839
xkb: release XTEST pointer buttons on physical releases. (#28808)
revealed a bug with the XTEST/PointerKeys interaction.
Events resulting from PointerKeys are injected into the event processing
stream, not appended to the event queue. The events generated for the fake
button press include a DeviceChangedEvent (DCE), a raw button event and the
button event itself. The DCE causes the master to switch classes to the
attached XTEST pointer device.
Once the fake button is processed, normal event processing continues with
events in the EQ. The master still contains the XTEST classes, causing some
events to be dropped if e.g. the number of valuators of the event in the
queue exceeds the XTEST device's number of valuators.
Example: the EQ contains the following events, processed one-by-one, left to
right.
[DCE (dev)][Btn down][Btn up][Motion][Motion][...]
^ XkbFakeDeviceButton injects [DCE (XTEST)][Btn up]
Thus the event sequence processed looks like this:
[DCE (dev)][Btn down][Btn up][DCE (XTEST)][Btn up][Motion][Motion][...]
The first DCE causes the master to switch to the device. The button up event
injects a DCE to the XTEST device, causing the following Motion events to be
processed with the master still being on XTEST classes.
This patch post-fixes the injected event sequence with a DCE to restore the
classes of the original slave device, resulting in an event sequence like
this:
[DCE (dev)][Btn down][Btn up][DCE (XTEST)][Btn up][DCE (dev)][Motion][Motion]
Note that this is a simplified description. The event sequence injected by
the PointerKeys code is injected for the master device only and the matching
slave device that caused the injection has already finished processing on
the slave. Furthermore, the injection happens as part of the the XKB layer,
before the unwrapping of the processInputProc takes us into the DIX where
the DCE is actually handled.
Bug reproducible with a device that reports more than 2 valuators. Simply
cause button releases on the device and wait for a "too many valuators"
warning message.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Acked-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Keith Packard <keithp@keithp.com>
When doing Xinerama, we'll dispatch font ops across all backend screens.
If using a font server (such that some operations can sleep), we'll put
the client to sleep once for each screen, but only wake up once, because
we're trying to keep track of the sleep count in _each_ screen's
closure.
Instead, just ask the core whether the client is already asleep.
Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Same as the matching key functions. Buttons, like keys, can have two states
for down/up - one posted, one processed. Posted is set during event
generation (usually in the signal handler). Processed is set during event
processing when the event queue is emptied and events are being delivered to
the client.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Support for the blend mode operators was added in
0ce42adbf4
and the requirement was bumped but when things were split off into
include/protocol-versions.h it defined it to 10. render uses
the lower of the client and server advertised versions so it's not
using the new blend mode operators.
Signed-off-by: Robert Hooker <sarvatt@ubuntu.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
When this privates.h is included in C++ builds, the compiler
complains about implicitly casting void* to void**. This small
patch fixes that up.
Signed-off-by: James Jones <jajones@nvidia.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
If a button release event is posted for the MD pointer, post a release event
through the matching XTEST device. This way, a client who posts a button
press through the XTEST extension cannot inadvertedly lock the button.
This behaviour is required for historical reasons, until server 1.7 the core
pointer would release a button press on physical events, regardless of the
XTEST state. Clients seem to rely on this behaviour, causing seemingly stuck
grabs.
The merged behaviour is kept for multiple keyboard PointerKey events, if two
physical keyboards hold the button down as a result of PointerKey actions,
the button is not released until the last keyboard releases the button.
X.Org Bug 28808 <http://bugs.freedesktop.org/show_bug.cgi?id=28808>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Problem:
lockedPtrButtons keeps the state of the buttons locked by a PointerKeys button
press. Unconditionally clearing the bits may cause stuck buttons in this
sequence of events:
1. type Shift + NumLock to enable PointerKeys
2. type 0/Ins on keypad to emulate Button 1 press
→ button1 press event to client
3. press and release button 1 on physical mouse
→ button1 release event to client
Button 1 on the MD is now stuck and cannot be released.
Cause:
XKB PointerKeys button events are posted through the XTEST pointer device.
Once a press is generated, the XTEST device's button is down. The DIX merges
the button state of all attached SDs, hence the MD will have a button down
while the XTEST device has a button down.
PointerKey button events are only generated on the master device to avoid
duplicate events (see XkbFakeDeviceButton()). If the MD has the
lockedPtrButtons bit cleared by a release event on a physical device, no
such event is generated when a keyboard device triggers the PointerKey
ButtonRelease trigger. Since the event - if generated - is posted through
the XTEST pointer device, lack of a generated ButtonRelease event on the
XTEST pointer device means the button is never released, resulting in the
stuck button observed above.
Solution:
This patch merges the MD's lockedPtrButtons with the one of all attached
slave devices on release events. Thus, as long as one attached keyboard has
a lockedPtrButtons bit set, this bit is kept in the MD. Once a PointerKey
button is released on all keyboards, the matching release event is emulated
from the MD through the XTEST pointer device, thus also releasing the button
in the DIX.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Problem:
lockedPtrButtons keeps the state of the buttons locked by a PointerKeys button
press. Unconditionally clearing the bits may cause stuck buttons in this
sequence of events:
1. type Shift + NumLock to enable PointerKeys
2. type 0/Ins on keypad to emulate Button 1 press
→ button1 press event to client
3. press and release button 1 on physical mouse
→ button1 release event to client
Button 1 on the MD is now stuck and cannot be released.
Cause:
XKB PointerKeys button events are posted through the XTEST pointer device.
Once a press is generated, the XTEST device's button is down. The DIX merges
the button state of all attached SDs, hence the MD will have a button down
while the XTEST device has a button down.
PointerKey button events are only generated on the master device to avoid
duplicate events (see XkbFakeDeviceButton()). If the MD has the
lockedPtrButtons bit cleared by a release event on a physical device, no
such event is generated when a keyboard device triggers the PointerKey
ButtonRelease trigger. Since the event - if generated - is posted through
the XTEST pointer device, lack of a generated ButtonRelease event on the
XTEST pointer device means the button is never released, resulting in the
stuck button observed above.
Solution:
This patch merges the MD's lockedPtrButtons with the one of all attached
slave devices on release events. Thus, as long as one attached keyboard has
a lockedPtrButtons bit set, this bit is kept in the MD. Once a PointerKey
button is released on all keyboards, the matching release event is emulated
from the MD through the XTEST pointer device, thus also releasing the button
in the DIX.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
When ResetCurrentRequest is called, or IgnoreClient is called when a
client has input pending, IgnoredClientsWithInput will be set. However,
a subsequent IgnoreClient request will clear the client fd from that fd
set, potentially causing the client to hang.
So add an Ignore/Attend count, and only apply the ignore logic on the
first ignore and the attend logic on the last attend. This is
consistent with the comments for these functions; callers must pair
them.
Fixes https://bugs.freedesktop.org/show_bug.cgi?id=27035.
Reviewed-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Keith Packard <keithp@keithp.com>
X server doesn't need to understand fpe internals, so use
register_fpe_functions from libXfont.
It's required to get new version of libXfont, therefore adjust it to be passed
to autoconf.
Signed-off-by: Tiago Vignatti <tiago.vignatti@nokia.com>
Reviewed-by: Mikhail Gusarov <dottedmag@dottedmag.net>
Reviewed-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Can't use next as a macro argument since we're accessing the .next field
of struct list.
Signed-off-by: Kristian Høgsberg <krh@bitplanet.net>
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Sometimes the vendor and product names aren't specific enough to target
a USB device, so expose the numeric codes in the ID. A MatchUSBID entry
has been added that supports shell pattern matching when fnmatch(3) is
available. For example:
MatchUSBID "046d:*"
The IDs are stored in lowercase hex separated by a ':' like "lsusb" or
"lspci -n".
Signed-off-by: Dan Nicholson <dbn.lists@gmail.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Serial input devices lack properties such as product or vendor name. This
makes matching InputClass sections difficult. Add a MatchPnPID entry to
test against the PnP ID of the device. The entry supports a shell pattern
match on platforms that support fnmatch(3). For example:
MatchPnPID "WACf*"
A match type for non-path pattern matching, match_pattern, has been added.
The difference between this and match_path_pattern is the FNM_PATHNAME
flag in fnmatch(3).
Signed-off-by: Dan Nicholson <dbn.lists@gmail.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
This patch was generated by the following Perl code:
perl -i -pe 's/([^_])return\s*\(\s*([^(]+?)\s*\)s*;(\s+(\n))?/$1return $2;$4/g;'
Signed-off-by: Mikhail Gusarov <dottedmag@dottedmag.net>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Signed-off-by: Keith Packard <keithp@keithp.com>
Allow InputClass sections to match against the running operating system
to narrow the application of rules. An example where this could be used
is to specify that the default input driver on Linux is evdev while it's
mouse/kbd everywhere else.
The operating system name is the same as `uname -s`, and matching is
case-insensitive.
Signed-off-by: Dan Nicholson <dbn.lists@gmail.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
The internals of XTest are used by Xi and Xkb, and both Xi and Xkb are
always required, so it makes little sense to have XTest place data in
a devPrivate, especially a devPrivate which is only available when the
XTest extension is enabled.
Signed-off-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Make sure all of the private keys used by the test code are
initialized before being used.
Signed-off-by: Keith Packard <keithp@keithp.com>
Tested-by: Robert Hooker <sarvatt@ubuntu.com>
Keys need to persist through server reset so that the private system
can be cleaned up in dixResetPrivates. In particular, this means that
keys cannot live in objects freed at reset time. This API provides
suitable object lifetime by having the privates code free the key in
the reset path.
Signed-off-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Each key now declares which object type it belongs to, this permits
keys for different types to share the same offset within the allocated
privates. As a special case for XSELinux, a key may be allocated as
PRIVATE_XSELINUX which will allow it to continue to be used across the
relevant object types.
Signed-off-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
This patch only changes the API, not the implementation of the
devPrivates infrastructure. This will permit a new devPrivates
implementation to be layed into the server without requiring
simultaneous changes in every devPrivates user.
Signed-off-by: Keith Packard <keithp@keithp.com>
Tested-by: Tiago Vignatti <tiago.vignatti@nokia.com>
For predefined resource types, the offset of the devPrivates field was
already kept in a constant table. The only non-predefined type needing
this treatment was dbeDrawableResType, which is just a magic alias for
RT_PIXMAP.
This patch special-cases looking up RC_DRAWABLE offsets and uses the
table directly for everything else.
Signed-off-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
This makes the region code changes source compatible with existing
code, although none of them are used within the server source itself.
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
This makes all of the previous macros into inline functions and also
turns all of the direct calls to pixman region code into inline
functions as well.
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
This is a combination of a huge mechanical patch and a few small
fixups required to finish the job. They were reviewed separately, but
because the server does not build without both pieces, I've merged
them together at this time.
The mechanical changes were performed by running the included
'fix-region' script over the whole tree:
$ git ls-files | grep -v '^fix-' | xargs ./fix-region
And then, the white space errors in the resulting patch were fixed
using the provided fix-patch-whitespace script.
$ sh ./fix-patch-whitespace
Thanks to Jamey Sharp for the mighty fine sed-generating sed script.
The hand-done changes involve removing functions from dix/region.c
that duplicate inline functions in include/regionstr.h, along with
their declarations in regionstr.h, mi.h and mispans.h.
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
This prepares the file to be moved from mi to dix. This patch
was done mechanically with the included scripts 'fix-miregion' run over
the entire X server and 'fix-miregion-private' run over
include/regionstr.h and mi/miregion.c.
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
Since reallocating the backing pixmap can fail, we need to try and do
it before any other side effects of reconfiguring the window happen.
This changes the ConfigNotify hook to return status, and moves the
composite window reconfiguration wrappers to ConfigNotify. They all
basically did the same thing, so we can drop the MoveWindow,
ResizeWindow, ChangeBorderWidth wrappers, and allow ConfigNotify to do
all the work. If reallocation fails we fail before we send any
confiureNotify events, or enter the area we can't recover from.
The only place we now enforce 32k limits are in EXA/UXA/fb, so drivers
that don't use this should probably deal with it in their pixmap
allocate if they don't already.
This also breaks ABI, so we need an alternate fix for older servers,
working on the X server makes me realise why I'm a kernel hacker.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Many references to the dixScreenOrigins array already had the
corresponding screen pointer handy, which meant they usually looked like
"dixScreenOrigins[pScreen->myNum]". Adding a field to ScreenRec instead
of keeping this information in a parallel array simplifies those
expressions, and eliminates a MAXSCREENS-sized array.
Since dix declared the dixScreenOrigins array, I figure allocating a
screen private for these values is overkill.
Signed-off-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Tiago Vignatti <tiago.vignatti@nokia.com>
Tested-by: Tiago Vignatti <tiago.vignatti@nokia.com> (i686 GNU/Linux)
Many references to the WindowTable array already had the corresponding
screen pointer handy, which meant they usually looked like
"WindowTable[pScreen->myNum]". Adding a field to ScreenRec instead of
keeping this information in a parallel array simplifies those
expressions, and eliminates a MAXSCREENS-sized array.
Since dix uses this data, a screen private entry isn't appropriate.
xf86-video-dummy currently uses WindowTable, so it needs to be updated
to reflect this change.
Signed-off-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Tiago Vignatti <tiago.vignatti@nokia.com>
Tested-by: Tiago Vignatti <tiago.vignatti@nokia.com> (i686 GNU/Linux)
Most references to the savedScreenInfo array already had the
corresponding screen pointer handy, which meant they usually looked like
"savedScreenInfo[pScreen->myNum]". Adding a field to ScreenRec instead
of keeping this information in a parallel array simplifies those
expressions, and eliminates a MAXSCREENS-sized array.
Since dix uses this data, a screen private entry isn't appropriate.
Signed-off-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Tiago Vignatti <tiago.vignatti@nokia.com>
Tested-by: Tiago Vignatti <tiago.vignatti@nokia.com> (i686 GNU/Linux)
Some drivers use DRI protocol but implement their own kernel rendering
manager. For these drivers, libdrm becomes useless. --disable-libdrm
configure parameter can be used to disable libdrm support in dri2.
To provide ABI/API compatibility for libdrm based drivers, libdrm call
is wrapped in ifdef.
Signed-off-by: Tiago Vignatti <tiago.vignatti@nokia.com>
Signed-off-by: Pauli Nieminen <ext-pauli.nieminen@nokia.com>
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Propagate the shape kind all the way to SetShape to avoid performing non-input
operations such as revalidating the tree and generating exposures when only
changing a window's input shape.
Signed-off-by: Pierre-Loup A. Griffais <pgriffais@nvidia.com>
Acked-by: Aaron Plattner<aplattner@nvidia.com>
Reviewed-by: Daniel Stone<daniel@fooishbar.org>
Signed-off-by: Keith Packard <keithp@keithp.com>
For absolute input devices (E.G. touchscreens) in multi-head setups,
we need a way to bind the device to an randr output. This adds the
infrastructure to the server to allow us to do so.
positionSprite() scales input coordinates to the dimensions of the shared
(total) screen frame buffer, so to restrict motion to an output we need to
scale/rotate/translate device coordinates to a subset of the frame buffer
before passing them on to positionSprite.
This is done here using a 3x3 transformation matrix, which is applied to
the device coordinates using homogeneous coordinates, E.G.:
[ c0 c1 c2 ] [ x ]
[ c3 c4 c5 ] * [ y ]
[ c6 c7 c8 ] [ 1 ]
Notice: As input devices have varying input ranges, the coordinates are
first scaled to the [0..1] range for generality, and afterwards scaled
back up.
E.G. for a dual head setup (using same resolution) next to each other, you
would want to scale the X coordinates of the touchscreen connected to the
both heads by 50%, and translate (offset) the coordinates of the rightmost
head by 50%, or in matrix form:
left: right:
[ 0.5 0 0 ] [ 0.5 0 0.5 ]
[ 0 1 0 ] [ 0 1 0 ]
[ 0 0 1 ] [ 0 0 0 ]
Which can be done using xinput:
xinput set-prop <left> --type=float "Coordinate Transformation Matrix" \
0.5 0 0 0 1 0 0 0 1
xinput set-prop <right> --type=float "Coordinate Transformation Matrix" \
0.5 0 0.5 0 1 0 0 0 1
Likewise more complication setups involving more heads, rotation or
different resolution can be handled.
Signed-off-by: Peter Korsgaard <peter.korsgaard@barco.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Peter wants to get a larger patch sequence put together and I didn't
read past the commit message to see the 'don't take this patch
please'.
This reverts commit 531ff40301.
Some input drivers need to implement an internal hotplugging scheme for
dependent devices to provide multiple X devices off one kernel device file.
Such dependent devices can be added with NewInputDeviceRequest() but they are
not removed when the config backend calls DeleteInputDeviceRequest(),
leaving the original device to clean up.
Example of the wacom driver:
config/udev calls NewInputDeviceRequest("stylus")
wacom PreInit calls
NewInputDeviceRequest("eraser")
NewInputDeviceRequest("pad")
NewInputDeviceRequest("cursor")
PreInit finishes.
When the device is removed, the config backend only calls
DeleteInputDeviceRequest for "stylus". The driver needs to call
DeleteInputDeviceRequest for the dependent devices eraser, pad and cursor to
clean up properly.
However, when the server terminates, DeleteInputDeviceRequest is called for
all devices - the driver must not remove the dependent devices to avoid
double-frees. There is no method for the driver to detect why a device is
being removed, leading to elaborate guesswork and some amount of wishful
thinking.
Though the input driver's UnInit already supports flags, they are unused.
This patch uses the flags to supply information where the
DeleteInputDeviceRequest request originates from, allowing a driver to
selectively call DeleteInputDeviceRequest when necessary.
Also bumps XINPUT ABI.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Dan Nicholson <dbn.lists@gmail.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
No special memory handling is used to give drivers the maximum flexibility
with the data. Drivers should be able to call realloc on the product string
if needed and perform similar operations.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Dan Nicholson <dbn.lists@gmail.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
For AtomError, this should fix a clang warning; in the other cases it's
just good documentation.
Signed-off-by: Jamey Sharp <jamey@minilop.net>
Cc: Jeremy Huddleston <jeremyhu@apple.com>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
This eliminates a poorly-named, poorly-documented field from the
ScreenRec, using a previously-unused flag bit in each GC instead.
Signed-off-by: Jamey Sharp <jamey@minilop.net>
Cc: Keith Packard <keithp@keithp.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
The structure containing the state of animated cursor was amended within
SpriteInfoRec, removing all previously privates logic to keep such state.
API change: It was removed MAXDEVICES dependency \o/
Signed-off-by: Tiago Vignatti <tiago.vignatti@nokia.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Keith Packard <keithp@keithp.com>
